Dans le cadre du cours INF6900A / INF7900, Pr. Omar Abdul Wahab organise une série de séminaires départementaux auxquels vous êtes tou(te)s convié(e)s. Veuillez trouver ci-dessous les détails du prochain séminaire, qui sera donné en anglais.
 

Conférencier : Daniel Pedronette / São Paulo State University (UNESP)

Titre : Rank-based Unsupervised Learning for Image Retrieva
 

Abstract: In several computer vision and retrieval applications, images are often represented by feature vectors and modeled as high-dimensional points in a Euclidean space. Comparisons are computed through a distance between their correspondent feature vectors, considering pairs of images and often measured by the Euclidean distance. However, the pairwise analysis is simplistic since it provides only locally restricted comparisons and ignores more global relationships and the dataset structure itself. More recently, research initiatives have focused on other stages of the retrieval process, which are not directly related to low-level feature extraction procedures. Methods have been proposed with the aim of ranking images with respect to the intrinsic global dataset structure, replacing pairwise distance with more global measures capable of considering the dataset manifold. This talk will focus on unsupervised learning methods based on ranking approaches for image retrieval tasks. In general, the main objective of such methods is to exploit the similarity information encoded on rankings aiming at improving the effectiveness of image retrieval tasks. Currently, other research directions have also been investigated, such as the use of the methods as a pre-processing step for classification tasks and the applications in other scenarios, such as multimedia and time-series retrieval.

Short-bio: Daniel Carlos Guimarães Pedronette received a B.Sc. in computer science (2005) from the State University of São Paulo (Brazil) and an M.Sc. degree in computer science (2008) from the University of Campinas (Brazil). He got his doctorate in computer science at the same university in 2012. He is currently an Associate Professor at the Department of Statistics, Applied Mathematics and Computing, State University of São Paulo (Brazil). His research interests involve content-based image retrieval, unsupervised and semi-supervised learning, image analysis, and information retrieval. He also serves as an Associate Editor of the Pattern Recognition journal.